Saurabh Garg - Mistral AI
Current Role
Saurabh Garg is currently involved in building next-generation AI models at Mistral AI. His work focuses on developing advanced artificial intelligence solutions.
Professional Experience and Background
- Mistral AI: Actively engaged in AI development.
- Apple: Completed an internship; notable work includes TiC-CLIP: Continual Training of CLIP Models, recognized at NeurIPS 2023.
- Google Brain: Internship focused on machine learning.
- Samsung Research Korea: Served as an engineer post-undergrad.
Education
- PhD in Machine Learning, Carnegie Mellon University (CMU): Saurabh completed his PhD under the guidance of Professors Zachary Lipton and Siva Balakrishnan. His work received prestigious fellowships from Bloomberg, JP Morgan, and Amazon.
- Undergraduate Degree from IIT Bombay, India: Graduated with honors in Computer Science and minors in Applied Statistics in 2018.
Collaborations and Fellowships
- Collaborations: Worked with renowned professors including Zico Kolter, Aditi Raghunathan, Hanie Sedghi, and Behnam Neyshabur during his tenure at CMU and internships.
- Fellowships:
- Bloomberg PhD Fellowship
- JP Morgan AI PhD Fellowship
- Amazon Graduate Fellowship
Notable Contributions and Research Highlights
Saurabh has made significant contributions to the field of AI and machine learning through numerous published works. Highlights include:
- 2023:
- Multiple research papers accepted at NeurIPS, including topics on contrastive learning, self-training, distribution shift, and optimal dynamic regret.
- Organized workshops on robustness in learning models at NeurIPS and ICML.
- 2022:
- Research on domain adaptation, unsupervised learning under label shifts, and characterizing data points accepted at NeurIPS.
- Received JP Morgan AI PhD Fellowship and Amazon Graduate Fellowship.
- 2021:
- Multiple presentations and accepted papers at ICML and ICLR on generalization bounds, gradient behavior, and learning from positive/unlabeled data.
- 2020 and earlier:
- Significant early contributions include work on neural architecture for question answering, generalization guarantees, and label shift estimation accepted at major conferences such as NeurIPs, ICML, and ECIR.
Professional Interests
Saurabh's professional focus lies in:
- Distribution Shift Management: Investigating methods to predict target domain performance and generalization guarantees under distribution shift.
- Neural Architectures: Developing frameworks for question answering and code-mixed speech recognition.
- Robust Machine Learning Models: Enhancing robustness and efficiency of AI models in various domains.
Online Presence and Additional Information
- LinkedIn: [Saurabh Garg - LinkedIn](https://www.linkedin.com/in/saurabh-garg-b680b5b8)
- Personal Website: [Saurabh Garg - Homepage](https://saurabhgarg1996.github.io/)
- Twitter: [@saurabh_garg67](https://twitter.com/saurabh_garg67)
Saurabh maintains an active professional presence, frequently updating his latest research, publications, and insights on platforms such as LinkedIn, his personal website, and Twitter.
This report provides a comprehensive view of Saurabh Garg’s professional journey, achievements, and current focus areas, useful for understanding his potential needs and contributions in the AI development landscape at Mistral AI.
